The Future (R)evolution of Preimplantation Genetic Diagnosis/Human Leukocyte Antigen Testing: Ethical Reflections


  • Guido de Wert Ph.D.,

    Corresponding author
    1. Research Institute Growth and Development, Faculty of Health, Medicine and Life Sciences, Section Health, Ethics and Society, Maastricht University, Maastricht, The Netherlands
    • Faculty of Health, Medicine and Life Sciences, Research Institute Growth & Development, Maastricht University, P.O. Box 616, 6200 MD Maastricht, The Netherlands. Telephone: 0031.433882145; Fax: 0031.433884171
    Search for more papers by this author
  • Inge Liebaers,

    1. Research Centre Reproduction and Genetics, Universitair Ziekenhuis Brussel, Brussels, Belgium
    Search for more papers by this author
  • Hilde Van de Velde

    1. Research Centre Reproduction and Genetics, Universitair Ziekenhuis Brussel, Brussels, Belgium
    Search for more papers by this author


There has been increasing support for combining preimplantation genetic diagnosis (PGD) for specific diseases with a test for human leukocyte antigens (HLA) because the generation of HLA-matched umbilical cord blood cells may save the life of a diseased sibling. To date, this procedure has taken place in the context of conceiving another child—PGD/HLA testing type 1. However, it may well become possible to perform PGD/HLA testing outside this context, that is, to select matched embryos from which embryonic stem cells could be derived and used in cell therapy—PGD/HLA testing type 2. A proactive ethical analysis is needed and is presented in this article. Although PGD/HLA testing type 1 can be morally justified, the risks, pitfalls, and practical limitations of this procedure make it necessary to develop alternative strategies. PGD/HLA testing type 2 may provide an alternative strategy. From an ethical point of view, the controversial issue is that this procedure creates embryos purely for instrumental use. However, given the dominant view that the preimplantation embryo has only limited moral value, this alternative may be as morally justified as PGD/HLA testing type 1.

Disclosure of potential conflicts of interest is found at the end of this article.


Recently, there was a strong consensus that preimplantation genetic diagnosis (PGD), like prenatal diagnosis (PD), should be restricted to the detection of mutations/chromosomal abnormalities that (may) affect the health of the particular prospective child or adult in questiona view called the medical model. The increasing support for combining PGD for a particular disease with a test for human leukocyte antigens (PGD/HLA testing) illustrates that a more permissive interpretation of this guiding principle is emerging, allowing preimplantation testing and selection for genetic characteristics that (may) affect the health of “third parties,” more specifically a diseased sib [1]. PGD/HLA testing currently takes place in the context of conceiving another child, sometimes called “parity for donation”; parity is the medical term used to refer to a woman who has given birth [2]. We define this strategy here as PGD/HLA testing type 1. In the future, however, it may well become possible to use the same PGD procedure to select matched embryos that will not be transferred into the uterus but will be used to derive embryonic stem cells for cell therapy. We will define this as PGD/HLA testing type 2. PGD/HLA testing type 2 would represent a completely new step in the ongoing technical evolution [1].

Before PGD/HLA testing type 2 comes into clinical practice, a proactive ethical analysis is needed and presented in this article. We will focus on hematopoietic stem cell (HSC) transplantation, recall the rationale for the need of HLA-identical HSC donors, and summarize the technical advances in hematopoietic differentiation from ESC lines. In the ethical discussion, we will first review the possible hurdles, pitfalls, and moral problems of PGD/HLA testing type 1. Secondly, we will compare the type 1 and type 2 cases from an ethical standpoint. We argue that, although type 1 is morally acceptable, type 2 may be a good alternative option for parents.

New Technologies to Obtain HLA-Identical Hematopoietic Stem Cells

HSC transplantation is an accepted treatment for a number of malignant and nonmalignant diseases of the hematopoietic/lymphoid system [3]. The HSC must be derived from an HLA-compatible donor source, which is ideally an HLA-identical sibling. Depending on the indication for transplantation and the source of the stem cells, some HLA mismatch could or should be allowed. In practice, unrelated donors of closely matched peripheral blood- or marrow-derived HSC can be found for 50% of the patients. In the case of cord blood, less stringent compatibility criteria may be applied; in the case of adults, a sufficiently matched cord blood graft or an acceptable combination of several cord blood grafts can be found for most of the patients. Unrelated HSC transplantation, however, including cord blood transplantation, holds an increased risk of transplant-related morbidity and mortality compared with HLA-identical transplantation. This is particularly pertinent to nonmalignant diseases (e.g., β-thalassemia, sickle cell disease) where transplantation is not urgent, graft-versus-leukemia activity is not required, and other treatment modalities may temporarily offer relatively adequate, although not curative, alternatives. When a sufficiently compatible unrelated donor source cannot be found and/or when an HLA-identical sibling is the only or best treatment option, conceiving another child as a potential donor of HLA-identical HSC may be considered.

Planning an additional child with the intention and hope of using it as a “donor” of HSC for an affected sib has been done for years [2]. In these cases, parents have decided to have another child by natural conception, hoping that it would be a good match. More recently, it has become possible to determine the suitability of the fetus as a future donor in utero [4, 5] in order to abort the fetus when there was no HLA-match. The newest step in this evolution is PGD/HLA testing type 1 [6, [7], [8], [9], [10], [11], [12]–13]. This strategy has some clear advantages above natural conception. First, although in theory only 1/4 embryos are HLA-identical and, in the case of PGD/HLA, 3/16 embryos (for autosomal recessive inheritance) or 2/16 (for X-linked recessive inheritance) are both healthy and HLA-identical, the chance of conceiving a matched donor increases substantially compared with a spontaneous conception because only the healthy (in case of PGD) HLA-matched embryos are transferred into the uterus. Second, selective transfer of a matched embryo avoids the problem of aborting a nonmatched fetus.

Pluripotent ESC may become a source of HSC in the near future. ESC can now be derived from the inner cell mass from human blastocysts with a reasonable efficiency [14, [15], [16], [17]–18]. Human ESC can also be induced to differentiate in vitro into multiple cell types including hematopoietic precursor cells [19, 20]. After further differentiation in vitro, the formation of erythrocytes, dendritic cells, B cells, natural killer cells, macrophages, and granulocytes has been reported [21, [22]–23]. These hematopoietic precursors are capable of long-term engraftment (homing) in a xenogeneic mouse transplantation model with no evidence of teratoma [24]. However, the capacity of the cells to maintain a completely functional hematopoietic-lymphoid system in patients remains to be demonstrated. In addition, the derivation and culture of human ESC lines in pathogen-free and animal-free conditions are a prerequisite for further therapeutic use, and several recent reports describe significant advances in this direction [25]. Considering this evolution in techniques, PGD/HLA testing type 2 may be a realistic alternative in the future because it would result in perfectly matched (healthy) ESC/HSC for transplantation.

The Ethics of PGD/HLA Testing Type 1


Three moral questions are identified. The first question is whether it is morally justified to use a child (an incompetent minor) who cannot decide for himself as a “donor” of a transplant. Clearly, the harm/probability ratio of the specific procedure(s) is of utmost importance. From a medical, moral, and legal perspective, the use of stem cells isolated from umbilical cord blood (usually 80–150 ml) is the simple case, as the cells are not part of the child and the procedure is noninvasive. Nevertheless, this procedure may raise moral questions. First, the timing of umbilical cord clamping may be important. Immediate or early cord clamping will probably best serve the interests of the recipient of HSC from the cord blood, as this method optimizes HSC collection. There are some recent claims, however, that delaying cord clamping (accompanied by lowering the infant to hasten the placental transfusion) offers protection to children, particularly for preterm and very low birth weight (VLBW) children [26, [27], [28]–29]. The finding that intraventricular hemorrhage and late-onset sepsis are more frequent in immediately clamped preterm infants is especially worrisome.

After early and, possibly, late clamping, the number of HSC may be insufficient to treat the diseased sibling properly. If this is the case, in vitro expansion of the HSC may become an option for some conditions, although this procedure needs to be further investigated [30, 31]. Today, a more pragmatic way to solve the problem is to opt for a repeat donation. The question is then whether the donor's bone marrow HSC can be collected. In young children, bone marrow aspiration, performed under general anesthesia, is necessary to obtain these HSC. Although this procedure is widely accepted, it is controversial. According to some ethicists, this can only be morally justified when there is evidence of a positive emotional relationship between the candidate donor and the potential recipient to ground an expectation of psychological benefit to the donor [32]. This strict view means that neonates and very young children should not be used as bone marrow donors.

The second question is whether it is acceptable to conceive a child (partly) in order to obtain cells for transplantation. Critics have various objections. The first objection concerns human dignity: the child would be used instrumentally, which is contrary to Kant's categorical imperative that “one should act in such a way that you always treat humanity, whether in your own person or in the person of another, never simply as a means, but always at the same time as an end” [33]. This criticism is, however, problematic. It is wrongly assumed that the only motive for having another baby is to obtain the required transplant material. Empirical research suggests that parents often have mixed motives for enlarging the family [34]. Furthermore, Kant's proscription is against using people solely as a means. What matters, then, is whether the parents will value the future child only as a transplant source or whether they will also love the child for itself [2]. The second objection concerns the psychosocial risks for all parties involved, particularly the well-being of the child born after HLA typing. The magnitude of this risk probably depends most on the quality of the relationships within the family; if the child feels that it is wanted just as much as the other children are, there is no reason to expect serious problems. It has been suggested that the child may feel empowered if the transplantation is successful but devalued if the sick child dies [35]. Follow-up studies will provide more information in the future. Another risk is that parents may feel pressure to opt for parity for donation. This could, clearly, endanger the welfare of both parents and children, especially if the parents cannot support more children for financial, medical, and/or psychosocial reasons. These pitfalls and risks underscore the importance of adequate psychological counseling to verify, as far as reasonably possible, the motives of the parents, their ability to fulfill additional parental responsibilities, and to assist them in making well-considered procreative decisions. In individual cases, significant uncertainties may remain around these issues.

Apart from the concerns above, it is important to realize that parity for donation carries substantial practical hurdles. Most importantly, the procedure is time consuming; even if the woman becomes pregnant immediately, there is a real risk that the diseased sib will die during the nine months before the donor is born. It is possible that, in the case of fast progressing diseases, parents (and doctors) might even be tempted to induce labor early in order to save the sib. Needless to say, this would be morally problematic, if not completely unacceptable, since it could seriously harm the donor child.

The third and most pertinent question is whether it can be morally justified to select an HLA-matched embryo for transfer in order to ensure that suitable transplant material may be acquired after birth. The objection of critics is that PGD/HLA testing is contrary to the medical model since the HLA type has nothing to do with the health of the future child. A first argument for adhering to the medical model strictly is the wedge or “slippery slope” argument: “Beware of designer babies!” This is not convincing. Although in case of HLA testing type 1 the medical model in the strict sense is abandoned, this does not mean that the “designer model” is embraced. After all, the baby is not being genetically molded to fit the parental picture of the perfect child. In view of this, the premise that anyone who rejects the designer baby model must, for the sake of consistency, also reject PGD/HLA testing type 1 is debatable. Furthermore, the argument that this strategy inevitably leads to designer babies in the future also looks problematic, since it assumes a certain automatism. The second argument not to make an exception to the medical model is that this procedure implies a huge wastage of “normal,” healthy in vitro fertilization (IVF)-embryos; in most cases, both affected embryos and unaffected embryos that are unsuitable for transplantation purposes will not be transferred (although couples having moral objections against the destruction of their healthy embryos may be offered to cryopreserve them in order to replace them later). This objection, however, is not convincing in view of both the dominant view that the moral status of preimplantation embryos is relatively low and the pending tragedy for the parents and the terminally ill child. Furthermore, most societies accept the loss of healthy embryos in regular IVF.

Which Conditions Should Be Imposed?

There is a growing support for the view that PGD/HLA testing type 1 may be morally justified [36]. The real issue, then, concerns the conditions that should be imposed. In the “standard” case, parents hope to save a sib affected with a hereditary condition and the HLA type selected for has no adverse consequences for the health of the future donor child. Three alternative scenarios complicate the ethical discussion. The first variant concerns the situation where the child suffers from a nonhereditary condition, like leukemia. There is dissent with regard to the ethics of PGD/HLA testing type 1 in this situation. Some consider this type of testing to be acceptable only in the context of IVF/PGD on genetic indication, that is, in addition to PGD of a genetic disease that may affect the future child. Outside of that context, so the argument runs, the possible health risks of IVF/PGD, especially of the biopsy, are disproportional. Recently, the Dutch Secretary of Public Health took this position [37]. The Human Fertilization and Embryology Authority, who originally imposed this restrictive policy in the United Kingdom, however, recently decided to opt for a more liberal guidance in view of the lack of evidence of harm caused by the biopsy procedure [38, [39], [40]–41].

The second and very exceptional variant would concern a diseased parent who is in need of HSC. Some commentators reject PGD/HLA testing type 1 in this situation because balanced parental decision making is compromised. They presumably fear that parents are not sufficiently able to consider the best interests of the donor child when their own lives are at stake. The parents may be more likely to choose another child while they are not able to take on additional parental responsibilities. This risk would, however, seem to be an additional reason for adequate counseling on a case-by-case basis rather than for a categorical rejection of the procedure.

Third, in exceptional cases, specific HLA types may carry substantial health risks for the future child. Several associations between various HLA types and specific disorders have been identified. One of the strongest associations is the one between the HLA-B27 polymorphism and Bechterew disease [42]. If Bechterew runs in the family, the carrier's risk may be as high as 20%. Would the current procedure be morally justified if the perfectly matched HLA type would impose substantially increased health risks to the future child [1]? A somewhat similar situation may occur in the case of insulin-dependent diabetes mellitus type 1, which is strongly associated (OR 21.5) with certain HLA class II alleles [43].

A Summary of the Obstacles

Clearly, PGD/HLA testing type 1 is a complex and demanding procedure for the parents. Pitfalls and problems are both practical and moral, and there is some overlap.

Practical problems/hurdles include: (a) Some parents cannot support another child for medical, economical, and/or psychosocial reasons; (b) The procedure has only a moderate take home baby rate (THBR) (10%) ([12, 13] and our unpublished data), and this is mainly due to the low number of embryos available for transfer. In the standard case, where PGD/HLA testing will be used to select embryos that are both free of a specific disease and a perfect HLA match, the chance of an embryo being both healthy and a suitable match is only 18.75% in the case of autosomal recessive conditions, such as β-thalassemia, and only 12.5% in case of sex selection for an X-linked disease; (c) Since the development of the HLA testing, establishing pregnancy, and pregnancy itself are time consuming, the diseased child will sometimes die before the HSC become available [34].

Ethically relevant risks for parents and donor child include the following: (a) Parents may feel under pressure to opt for this procedure while they cannot afford another child (“an offer you cannot refuse”); (b) The donor child might feel devalued (although clearly the opposite may be true); (c) Possible health risks for the future donor child are related to the blastomere biopsy (a risk so far considered to be low) [38, [39], [40]–41], specific HLA types that are associated with familial disorders [42, 43], and early umbilical cord clamping, which may have adverse effects to premature and/or VLBW babies [26, [27], [28]–29].

In view of these problems, it is necessary to develop possible alternative solutions. One future option might be PGD/HLA testing in order to select matched embryos from which ESC could be derived to produce HSC for cell therapy—PGD/HLA testing type 2.

The Ethics of PGD/HLA Testing Type 2

Research on the controlled differentiation of ESC, for example to HSC, is advancing rapidly [21, [22]–23]; therefore, a proactive debate on the practical and ethical pros and cons of the type 2 case is important. Can this procedure be morally justified and, if so, under which conditions?


The first advantage of PGD/HLA testing type 2 is that it will increase the number of therapeutic options. For those parents who cannot support another child, this procedure may be the only real option. Other parents, if given the choice, may prefer PGD/HLA testing type 2 to type 1, especially if they consider their family to be complete. For these parents, the type 1 case will be second best.

Second, PGD/HLA testing type 2 may have substantial medical/practical advantages if it were to have a higher success rate than the type 1 procedure in terms of saving the diseased sib. Clearly, this remains to be seen. One may need fewer embryos to obtain the cells needed for transplantation: the THBR after PGD/HLA is only 10% ([12, 13] and our unpublished data), whereas the derivation rate of a stable ESC line from an embryo after PGD is approximately 20% [18]. Furthermore, a successful pregnancy takes 9 months, whereas the type 2 procedure may be less time consuming; generally, it takes 3 months to derive and characterize a stable ESC line and 2–3 weeks to differentiate these cells into a heterogeneous population containing hematopoietic progenitor cells [19, 24]. However, before it can be used for HSC transplantation in patients, a high number of cells would be required, and the safety of the ex vivo expansion and the transplantation still need to be evaluated. Finally, the availability of HLA-matched ESC could perhaps make it possible in the long-term to derive other tissues/organs in case another transplantation is required (e.g., kidney).

Third, the child-related objections and concerns regarding type 1 would be circumvented. These include the presumed instrumental use of children and the possible psychosocial and medical risks for the future donor. As a consequence, the debate about the conditions to be imposed in view of the additional complications inherent in the variants mentioned above would become irrelevant. The cases of a diseased parent, a nonhereditary disorder, and possible additional tissue donation would become nonissues in the context of PGD/HLA testing type 2. Likewise, the selected HLA type would never imply a health risk for future donors.


The primary normative issue regarding PGD/HLA testing type 2 concerns the creation of (human) preimplantation embryos solely for so-called instrumental use, particularly, to obtain ESC/HSC for cell therapy. It is important to distinguish between a legal and an ethical perspective. Many countries prohibit the creation of embryos for instrumental use in accordance with Article 18 of the European Convention on Human Rights and Biomedicine of the Council of Europe: (1) Where the law allows research on embryos in vitro, it shall ensure adequate protection of the embryo; (2) The creation of human embryos for research purposes is prohibited [44]. At the same time, many countries accept the instrumental use of so-called surplus or spare embryos, which remain after IVF or IVF/PGD. From an ethical perspective, the question arises as to whether there is a decisive moral difference between the instrumental use of spare embryos and the creation of embryos for instrumental use [45]. Why accept the former and categorically prohibit the latter? The ethical and societal debate about this issue often focuses exclusively on the moral status and the protection of the embryo. It is important, however, not to treat embryos as entities disconnected from their source, as if the female body is not involved. After all, embryos come from oocytes, and oocytes come from women. The ethical debate about creating embryos for instrumental use should, therefore, consider both the moral status of the embryo and the interests and welfare of women, especially candidate oocyte donors. According to Raymond, the feminist perspective, which focuses on women/oocyte donors involved, is as important as the fetalist perspective, which focuses on the embryo [46].

From a fetalist perspective, there are good reasons to consider that there is no fundamental difference between the instrumental use of spare embryos and the creation of embryos for instrumental use, since the moral status of these embryos is the same. However, there is a difference with respect to the intention at fertilization. This difference, however, is only gradual; it is a misunderstanding to assume that in the context of regular IVF, each embryo is created as a goal in itself. The loss of some spare IVF embryos is a calculated risk beforehand. It can, therefore, be argued that the creation of embryos for instrumental use, just as the use of spare embryos, can be morally justified on grounds of proportionality (i.e., the instrumental use of embryos should serve an important goal or, more specifically, an important health interest) and subsidiarity or necessity (i.e., there is no suitable embryo-saving alternative to reach this goal). Assuming that the moral status of the early embryo (which will be used at day 5) is relatively low, the use of embryos for life-saving cell therapy is clearly proportional. Regulations in some countries (e.g., the U.K. and Belgium) permit somatic cell nuclear transfer for therapeutic purposes, so-called therapeutic cloning. It would be inconsistent, then, not to allow the instrumental use of embryos for other types of cell therapy, like in PGD/HLA testing type 2.

Critics of PGD/HLA testing type 2 would possibly argue that this strategy is contrary to the principle of subsidiarity, pointing to various alternative options that are embryo-saving. The first option would be to invest further in cell banks in order to maximize the HLA types available for treatment. It seems impossible, however, to guarantee that cell banks would be able to provide matched cells for all patients, irrespective of their HLA type [47]. Thus, PGD/HLA testing type 2 may still be needed as a last resort in individual cases. Another alternative, which limits the loss of embryos, is PGD/HLA testing type 1, as the matched and healthy embryos will be transferred. Can one indeed convincingly argue that PGD/HLA testing type 2 is morally unjustified if type 1 would be possible? This restrictive interpretation of the principle of subsidiarity is not reasonable in view of the possible advantages of type 2; this strategy may better meet the interest of both the diseased sib (as the procedure may have a better success rate) and the parents (they will have more options). Furthermore, type 2 avoids the psychological and medical risks of type 1 for the donor child. In other words, the restrictive interpretation of the principle of subsidiarity illustrates a one-dimensional approach that a priori gives priority to avoiding or minimizing loss of preimplantation embryos. A permissive interpretation of this principle, then, seems to be justified.

Obviously, it is important to acknowledge that the difference between PGD/HLA testing type 1 and type 2 is only gradual, as the loss of embryos can be 75%, 81.25%, or 87.5% in the first case (depending on whether the embryo is only tested for its HLA type, in combination with an autosomal recessive disorder, or an X-linked disorder by sexing, respectively) and 100% in the latter. The argument that the first is acceptable whereas the latter is not is difficult to accept.

From a feminist perspective, the production of embryos for instrumental use is controversial mainly because both the hormones given in order to induce a “super-ovulation” and the oocyte aspiration are burdensome and carry some risk for women (the oocyte donors). Critics consider this to be especially problematic when women are asked to donate oocytes for research purposes. They argue that the benefits of research are uncertain, if not speculative; the risks and burdens of the procedure for the donor are disproportional and, therefore, unjustified. However, this criticism seems to be less relevant for the ethics of oocyte donation for therapeutic purposes, as in PGD/HLA testing type 2. In this case, the risks and burdens of oocyte donation seem to be proportional in view of the benefit to be gained, that is, saving the woman's child. It would be inconsistent to accept PGD/HLA testing type 1 and, at the same time, to reject type 2 because of the risks/burdens of the latter to women involved. One may even argue that type 2 is less burdensome for women who would prefer not to have another pregnancy and child.


Considering the reflections above, the following conclusions may be drawn: (a) The traditional guiding principle holds that PGD, like PD, should focus on the diagnosis of genetic defects that (may) affect the health of the particular potential child. There is (rightly) increasing support for a more permissive view that also allows preimplantation testing for embryonic characteristics that may be highly relevant for the health of third parties, particularly PGD/HLA testing; (b) The ethical debate has so far disregarded the possible dynamics of PGD/HLA testing. It is important to distinguish between (the presently performed) type 1 and (the possible alternative) type 2; (c) Although PGD/HLA testing type 1 can be morally justified, the risks, pitfalls, and practical limitations of this procedure make it imperative to develop alternative strategies; (d) One future alternative may be PGD/HLA testing type 2. From an ethical point of view, the sensitive point with this procedure is that it involves the creation of embryos purely for instrumental use. Considering the dominant view that the preimplantation embryo has only limited moral value, this alternative may be morally justified as well. Both a prohibition of the creation of embryos for instrumental use (either in research or in future therapy) and the restrictive interpretation of the principle of subsidiarity are questionable. If safe and effective, type 2 is, prima facie, even the better option, from a medical, psychosocial, and ethical point of view; (e) Adequate preclinical studies with embryos donated for research with respect to the efficiency and safety of PGD/HLA testing type 2 (including methods to expand the cells ex vivo prior to transplantation) are of utmost importance. Since developmental biology is evolving fast, these laboratory experiments and the ethical discussion should be conducted simultaneously.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.